The Work Ability Index (WAI) is a widely used tool that measures the perception of sustainability at work derived from the construct of WA, developed by Ilmarinen and the research group of Finnish Institute of Occupational Health in the 1980s. In the past, this construct was used mainly by occupational physicians, but it is now fully integrated in the occupational health psychology (OHP) literature (Brady et al., 2016).

WA describes the physical and intellectual resources on which individuals can rely to respond to the emotional, cognitive, and physical demands posed by their work (Tuomi et al., 1991, 2001). WA has a strong health-related nature; hence, it inevitably decreases with age, even if there is a high rate of variability among individuals, as widely demonstrated by international research (Ilmarinen et al., 1997; Pojonen, 2001; Goedhard and Goedhard, 2005; Camerino et al., 2006; Lin et al., 2006; Monteiro et al., 2006; van den Berg et al., 2008, 2009; Converso et al., 2015b; Godinho et al., 2016) that highlights, for example, the same trend within different European countries (Camerino et al., 2006). Ilmarinen and Ilmarinen (2015) explained that this decline is due to the imbalance between the individual resources and the demands of the job, namely the work and the organizational context are not aligned to the age-related changes in personal resources. Poor WA is predictive of sickness absences, early retirement, disability pension, intention to leave (Goedhard and Goedhard, 2005; Camerino et al., 2006; Monteiro et al., 2006; Emberland and Knardahl, 2015), work stress, depression (Tuomi et al., 1991; van den Berg et al., 2008; Bethge et al., 2009; Godinho et al., 2016), and emotional exhaustion (Viotti et al., 2017a). Conversely, good WA is associated with a high quality of work, high productivity, and enjoyment of time at work. WA also predicts good quality of life, well-being, and active and meaningful retirement (Tuomi et al., 2001).

A recent review of the WA model, appropriate for an aging workforce process, the WA-PR (Work Ability Personal Radar), assumes a multidimensional view of WA based on the self-assessment of subjective experiences of personal resources, working context, and work–life interface and proposes the metaphor of a house, whose floors represent the four main dimensions of WA (Ilmarinen et al., 2005, 2015). Health and Functional Capacity is the first floor of the house and consists of physical, mental, and social resources; Occupational Competence, the second floor, consists of expertise, skills, and knowledge acquired through experience, years of work, training, and education; Attitude and Motivation, situated on the third floor, are the reasons people work and are influenced by previous work experiences; Working Conditions correspond to the fourth floor and consist of work and all of its dimensions. If the resources of the individual (floors 1–3) are in balance with the fourth floor, the WA is good. If the individual resources are in imbalance with the working conditions, WA is deteriorating. Organizational support is of great importance in relation to the fourth floor, while the personal resources are of great importance in relation to the first three floors.

Altogether, the floors of the WA house are in relation to a context of life that consists of work, family, and spare-time activities, representing the fifth floor (Ilmarinen et al., 2015).

This new multidimensional model of WA overcomes the previous conceptualization, mainly focused on an individual state, and coherently with other OHP models (Job-Demand-Control, JD-C, of Karasek, 1979 and Job-Demand-Resources, JD-R, of Demerouti et al., 2001) consider the interaction between personal resources (the first three floors) and the job demands and resources (the fourth floor). As it is assumed by the JD-R model (Demerouti et al., 2001), work factors which could affect workers’ psychological health can be classified into two categories: job demands and job resources. Job demands refer to those aspects of job that require the workers to sustain psychological or physical effort and are associated with cost in terms of health. Job resources refer to those physical, psychological, social, or organizational aspects of the job that may help to achieve work goals, reduce job demands and the related physiological and psychological costs, and stimulate personal growth and development.

The structure of WA changes during a person’s life and career as, for example, aging affects the individual’s resources or personal resources like competence and experience that may completely or partially compensate for physical “losses” and changes (Baltes and Baltes, 1990). Moreover, when people perceive their time as limited (as older workers do), emotional goals take priority and workers tend to focus their attention on affective rewards and resources (Carstensen et al., 1999; Truxillo et al., 2012) rather than on material or career-based rewards. As Demerouti et al. (2001) have affirmed, each work environment includes different demands and resources that interact with personal resources (Xanthopoulou et al., 2007) developing a dual process, the health impairment process and the motivational process, which may vary throughout the lifespan, when the same job demand may be perceived as more demanding, or the same resource may be perceived as less supportive. In other words, job well-being along the entire work lifespan is the result of an ever–changing balance between individual resources and job demands.

Several studies have shown that WA is influenced by individual, job-related, and lifestyle factors (Tuomi et al., 1998; Camerino et al., 2006; Bethge et al., 2009; Matzolumi et al., 2012; McGonagle et al., 2014, 2015; Airila et al., 2014; Converso et al., 2015a; Godinho et al., 2016; Li et al., 2016; Sottimano et al., 2017; Viotti et al., 2017b). To understand the factors that sustain WA, researchers have analyzed specifically the role of job resources in supporting WA, both directly (Tuomi et al., 1991; Larsson et al., 2012; Matzolumi et al., 2012; Airila et al., 2014; McGonagle et al., 2014; Emberland and Knardahl, 2015; Li et al., 2016) and indirectly, by buffering the negative effects of job demands on WA (McGonagle et al., 2014; Viotti et al., 2017b). Specifically, studies have highlighted the direct effect of decision authority, skill discretion, social support, and the meaning of work on WA in various working populations (van den Berg et al., 2008, 2011; Weigl et al., 2013; McGonagle et al., 2014; Li et al., 2016; Sottimano et al., 2017). For example, Li et al. (2016) showed that social support was positively and directly associated with WA, while Weigl et al. (2013) and McGonagle et al. (2015) underlined the importance of control in maintaining WA. In particular, Weigl et al. (2013) have shown that job control moderates the negative impact of aging on WA. In this direction, in a study by McGonagle et al. (2014) among healthcare workers, supervisor support and skill discretion were found to moderate the negative relationship between job demands and WA. Schulz et al. (2017) have highlighted similar findings: team health climate emerged as particularly important for older employees’ WA because the social context may support those compensation strategies that older workers may employ when job demands exceed their physical or mental capabilities.

Furthermore, the lifespan perspective on work design (Truxillo et al., 2012; Zacher and Schmitt, 2016) proposes that some job resources are more positively related to indicators of occupational well-being among older workers. More specifically, the literature shows that decision authority (Zacher and Frese, 2009; Zacher et al., 2010; Truxillo et al., 2012; Zaniboni et al., 2016; Viotti et al., 2017b), skill discretion (Truxillo et al., 2012; Zaniboni et al., 2013, 2014; Viotti et al., 2017b), social support (Truxillo et al., 2012; Drabe et al., 2014; Li et al., 2016), and meaning of work (Sottimano et al., 2017; Viotti et al., 2017b) are very important dimensions to promote WA and, more generally, well-being at work, in particular within older workers.

Personal resources are considered positive self-evaluations that are linked to resilience and refer to an individual’s sense of ability to successfully control or impact his or her environment (Hobfoll et al., 2003); they may have similar motivational potential to those of job resources (Xanthopoulou et al., 2007). For example, Youssef and Luthans (2007), in the psychological capital model perspective, found that hope had a positive effect on employee satisfaction and work happiness (Luthans, 2002; Luthans and Youssef, 2004); other researchers have shown that optimism correlates with employee engagement and performance (Medlin and Faulk, 2011), while resilience positively correlates with workplace performance and work happiness (Luthans et al., 2008).

Few studies have specifically considered the relationship between WA and personal resources. In this context, Airila et al. (2014) highlighted that personal resources predicted WA both directly and indirectly via work engagement, whereas the studies by Larsson et al. (2012), specifically conducted within nurses and nurse aides, and Ng et al. (2015) revealed that self-efficacy contributed to increased WA.

As previously stated, the topic WA changes across the lifespan and needs to be addressed differently according to the job characteristics, the job demands, and the organizational context. According to Warr (1994), four main types of work emerge from the interaction between age/experience and job demand: (a) the first one includes jobs in which the required skills do not decline with age and experience is an added value to the work; (b) in the second one, the skills required do not decline with age, but experience does not constitute an added value to the work; (c) the third type includes workers whose skills decline with age, but for which experience can partially compensate; (d) in the fourth, the required skills decline with age and experience cannot adequately compensate. In this direction, the literature shows that the relationship between age and ability to work is more critical when the workers are engaged in jobs with high physical demands (McGonagle et al., 2015). Often, for these professions, an early retirement is provided or a transition to less-physically demanding job tasks is made available within the same organization. In addition, WA can be supported through prevention activities and radical job redesign interventions: a famous example of good practice is the BMW intervention developed in 2007 (Loch et al., 2010), where a relatively low-cost project mainly directed toward preventing musculoskeletal disorders in a high-physically demanding job where the workforce was significantly aging, increased the productivity by 7%. Introducing in the first experiment 70 ergonomic changes, like installing special chairs at several workstations, allowed workers to spend the majority of their working hours sitting down or relaxing for short periods; vertically adjustable tables to adapt the workstations to each worker’s height, and so reducing back strain; special lenses to help workers distinguish among small parts, reducing eyestrain and mistakes due to the decline of sight. As scholars affirmed, BMW “defused the demographic time bomb” redesigning its factory for and with older workers, who were involved asking them to imagine how to ameliorate their working conditions in light of the aging process (Loch et al., 2010).

When the job demands are mainly cognitive, as in most of the type A or B jobs, WA decreases slowly and a large number of interventions devoted to sustainability in the organizational context can be developed. Regarding emotionally demanding jobs, they may be included in types A or C as, for example, job experience can be helpful in contrasting customer-related stress. One specific case is represented by jobs having both high physical and emotional demands, as in healthcare and pre-school contexts, where job redesign interventions or the adaptation of work environment are more complex or impossible to promote because of the specific nature of the work. Taking care of patients, babies, children, and sick people represent the core activity of these jobs that cannot be reduced or changed, and the continuous exposure to bio-mechanical risks may increase musculoskeletal disorders (Phongamwong and Deema, 2015), especially during aging, and negatively impact WA (Prochnow et al., 2013). As highlighted by previous studies conducted among preschool teachers aged 50 or over, the perception of poor WA (associated with physical problems) can mediate the relationship between demands/resources and worker well-being (Viotti et al., 2017a).

In light of the above, the present study aims then at examining the role of job and personal resources between age and WA within a typical “high touch” profession: nurses. We considered as job resources three dimensions on the basis of the previously described WA-PR house model and of the studies on the relationship between job resources and WA (Weigl et al., 2013; McGonagle et al., 2015; Li et al., 2016): decision authority that describes the autonomy in making job-related decisions (as autonomy), skill discretion (Viotti and Converso, 2016) that describes the opportunity to use certain skills at work (as competence), and social support (Li et al., 2016), as WA in the nursing context can be supported by job resources of an interpersonal nature, such as social support (Hägglund et al., 2011; Airila et al., 2014) expressed by colleagues, superior and patient as well (Converso et al., 2015a); meaning of work that can be considered as a type of intrinsic reward referred to job meaningfulness and constructiveness (Viotti et al., 2017b), for its importance in the field of the “helping professions.”

Concerning personal resources, we considered the following factors: self-efficacy, hope, optimism, and resilience. Self-efficacy refers to people’s beliefs in their capabilities to produce desired effects by their own actions (Bandura, 1977); hope refers to a positive motivational state; optimism refers to the tendency to expect positive events in one’s life; and resiliency refers to successfully coping with adversity or stress.

We hypothesize the following:

The study was cross-sectional and non-randomized. Data were collected by means of a self-report questionnaire during a multi-center survey conducted in two hospitals in a region of North–West Italy in 2016. Questionnaires were distributed during working hours. Workers were asked to enclose the completed questionnaire in an envelope and leave it in a box that the research team had placed in each unit/service. In total, 524 nurses were reached and 333 questionnaires were returned to the research team (response rate: 63.54%). The majority were women (female = 85.9%; male = 14.1%), with an age ranging from 23 to 64 years (M = 44.65, SD = 16.02). The average job seniority in the health sector was 19.53 years (SD = 10.61) and ranged from less than 1 to 42 years. Finally, 64% of the sample worked on the night shift.

The distribution of the socio-demographic data in this sample is analogous to the distribution in the Italian nursing workforce employed in the public health sector in which 65% are women and 35% are men, the average age is 44.6, and the average job seniority is 17 years¹.

Hospital administrations evaluated, endorsed, and authorized the research, allowing researchers to use the data for scientific purposes. Upon approval, department chiefs and nursing coordinators from each unit/service were asked for authorization to administer the questionnaire to the nurses. An additional ethical approval was not required since there was no treatment, including medical, invasive diagnostics, or procedures, causing psychological or social discomfort for the participants, nor were patients the subject of data collection.

Participation in the survey was voluntary. The research conforms to the provisions of the Declaration of Helsinki in 1995 (as revised in Edinburgh 2000), and all ethical guidelines were followed as required for conducting human research, including adherence to the legal requirements of the study countries.

The questionnaire was developed specifically for this study. It included items aimed at collecting age, other socio-demographic information (i.e., gender, job seniority, and night shift) and sub-scales aimed at measuring study variables (i.e., work ability, job resources, personal resources, and physical demands).

Psychometric proprieties of the measures used were found to be adequate. Regarding work ability, CFA showed an excellent fit, after adding two covariances between two couples of error terms: χ² = 35.86, df = 12; χ²/df = 2.98; CFI = 0.92; GFI = 0.97; SRMR = 0.05; RMSEA = 0.07(0.05–0.10). The second-order CFA including all personal resources revealed an adequate structure, after incorporating four covariances between error terms (i.e., three covariances between error terms of four observed variables loaded on the resilience factor and one covariance between two error terms of two items leaded on the optimism factor): χ² = 512.93, df = 267; χ²/df = 1.92; CFI = 0.91; GFI = 0.90; SRMR = 0.05; RMSEA = 0.05(0.05–0.06). Finally, the second-order CFA for job resources showed an excellent fit as well: χ² = 344.20, df = 147; χ²/df = 2.34; CFI = 0.92; GFI = 0.90; SRMR = 0.05; RMSEA = 0.06(0.05–0.07).

In all the CFAs, all the observed variables significantly loaded on its corresponding latent factor and in the expected direction. Moreover, as regarded the second-order CFA, all the first-order factors positively and significantly loaded on the second-order factor.

Discriminant validity between WA and each resource of personal and job type was proved. As shown in Table 1, in all cases, the two-factor model, where each item leaded on its corresponding factor (Table 1, a column), fitted significantly better than the one-factor model, where items from work ability and a resource considered loaded on the same factor (Table 1, b column).

Table 2 shows the univariate relationships between the variables under study, including control variables. As expected, WA showed a significant and negative correlation with age. Moreover, WA was found to be positively associated with all resources considered, with the exception of support from superiors. Regarding control variables, WA was significantly and negatively associated with job seniority. Finally, t-test showed that those who worked on the night shift reported significantly higher levels of WA (M = 39.27; SD = 5.38) when compared with those who did not work on the night shift (M = 37.38; SD = 5.80).

Table 3 shows the results of moderated regressions assessing the moderating role of job resources in the relationship between age and WA. In the first step, in all models, R² was significant, and age was significantly and negatively associated with WA; moreover, all the resources considered were found to be associated with WA, except for support from superiors. In the second step, the interaction effect between age and job resources was found to be significant in two of the five models carried out, suggesting that decision authority (β = 0.16, p = 0.003) and meaning of work (β = 0.14, p = 0.011) moderate the effects of age on WA. In the third step, when the models were adjusted for the control variables, both of the interaction terms kept showing significant β values (i.e., age ^∗ decision authority: β = 0.16, p = 0.004; age ^∗ meaning of the work: β = 0.14, p = 0.014). Regarding control variables, job seniority was found to be positively associated with WA (β = 0.22, p = 0.017) in one model only (JR4: support from superiors).

Slope test analyses were performed in order to further examine the direction of the effects of decision authority and meaning of work in the relationship between age and WA. As regards decision authority (Figure 1), simple slope analysis showed that when this resource was high (+1 standard deviation, SD), the relationship between age and WA (t = -2.28, p = 0.023), albeit significant, was weaker if compared with a condition of low job resource (-1 SD, t = -5.93, p = 0.001). Similarly, (Figure 2), the relationship between age and WA was stronger in the condition of low meaning of work (-1 SD, t = -5.24, p = 0.001) rather than in the condition of a high meaning of work (+1 SD, t = -3.01, p = 0.03).

Table 4 shows the moderated regressions assessing the moderating role of personal resources in the relationship between age and WA. Both in the first and in the second step, all the models reported a significant R², age was found to be significant in all models, and all personal resources considered were found be positively associated with WA. Moreover, in two models, the interaction term was found to be significant at both steps 1 and 2, supporting the moderating role for hope (step 1: β = 0.14, p = 0.011; step 2: β = 0.12, p = 0.027) and resilience (step 1: β = 0.11, p = 0.022; step 2: β = 0.12, p = 0.029), but not for optimism and self-efficacy, in the relationship between age and WA. As regards control variables (step 3), no significant relationships with WA were reported.

Simple slope analysis (Figures 3, 4) showed the relationship between age and WA was stronger in conditions of low personal resources (-1 SD, hope: t = -7.34, p = 0.001; resilience: t = -5.02, p = 0.001) as opposed to conditions of high personal resources (+1 SD, hope: t = -1.42, p = 0.15; resilience: t = -2.11, p = 0.035).